Critical Thinking Credit:medium.com
Introduction
As AI is increasingly used to solve complex problems, critical thinking has become a key component for its successful implementation and value generation. Educated individuals with sharpened critical thinking skills are not just capable of solving problems AI has difficulty with. In addition, they are also better positioned to successfully leverage AI, improving its own capabilities and weeding out bad actors. No clever prompt typed into an AI tool will ever be able to replace human critical thinking, especially as tools like ChatGPT often provide wrong results. Hence, as AI takes on the role of a thinking machine, applying critical thinking is crucial.
Historic Overview
The history of language models starts in 1883 with a concept of semantics, developed by the French philologist, Michel Bréal. He studied the ways languages are organised and how words are connected within a language. Based on his work, Natural Language Processing (NLP) gained great popularity following the end of World War II in 1945. Negotiating peace, individuals responsible for the outcome realised that translating from one language to another was crucial. In 1950 Alan Turing published a simple method of determining whether a machine can demonstrate human intelligence: If a machine can engage in a conversation with a human without being detected as a machine, it has demonstrated human intelligence. In 1958, Frank Rosenblatt created the first artificial neural network (ANN), known as the Mark 1 Perceptron. Released in 1966, ELIZA was the first computer program able to conduct human-like conversations. It could recognise simple user inputs and respond based on predefined scripts that gave users the illusion of understanding by showing emotions and empathy. In the early 90s text analysis and speech generation with Recurrent Neural Networks became very popular. In 2006, Google Translate was launched as a multilingual neural machine translation service. It was able to translate text, documents and websites from one language to another. Apple’s Siri, released in 2011 was the first successful NLP/ AI assistant. Siri’s automated speech recognition module translates the user’s words into digitally interpreted concepts. The voice-command system then correlates those concepts to predefined commands and performs specific actions. In 2017, the Transformer architecture was introduced by researchers working at Google. This model incorporates mechanisms to capture dependencies and relationships within input sequences and to improve previous architectures for machine translations. The Transformer model led to the development of Generative Pre-trained Transformers (GPT) and Bidirectional Encoder Representations from Transformers (BERT). In 2020 GPT-3 provided another big performance improvement over the previously released GPT-2. In 2022, OpenAI released ChatGPT which significantly surpasses GPT-3 in a range of tasks, including communicating in human-like English, developing new software, and writing speeches. There is no end in sight, as to how further improvements will increase the capabilities of learning machines.
What is Critical Thinking?
Critical thinking skills are also defined as ‘higher order’ skills which require ways of thinking that are deeper and more complex than the kind of ‘everyday’ thinking that we use to cook a meal, for example. A well-known framework that describes different levels of thinking is Bloom’s Taxonomy of Educational Objectives (1956). This framework suggests that remembering, understanding, and applying facts, figures, concepts or other ways of learning are ‘lower order’ skills. In contrast, there are three ‘higher order’ skills that represent the core of critical thinking. They consist of the following:
Analysis
This involves close reading or scrutiny of a piece of work to detect and identify its main points, arguments and conclusions as well as the evidence offered in support of them. Analysis often involves comparing and contrasting the work of different authors, identifying key themes or areas of contention, or making connections between different ideas or approaches towards the topic under consideration. Analysis may also involve the detailed examination of other data, such as the outcome of an experiment, a computer simulation or responses to a survey.
Evaluation
Evaluation involves assessing and probing the various points, arguments and evidence that one has found in order to make a judgement about their credibility, relevance and strength. It may involve considering the facts which have been omitted or included and questioning the conclusions that have been reached.
Inference
Inference involves reaching a conclusion based on the analysis and evaluation of the available information. This may involve agreeing or disagreeing with the theories, arguments and conclusions of others, discussing the implications of the information that one has considered and possibly making suggestions or recommendations for the future.
Critical Thinking Is the Key to AI
Generative AI has revolutionized AI development to connect GPT’s to a wide range of custom solutions. However, this progress demands a solid foundation of truthful data sourced from human experts. Human-derived insights are vital throughout production of large language models like ChatGPT in order to improve quality and fine-tune it for specific use cases. Consulting major global companies, Helen Lee Bouygues (age 53), one of the most successful women in business transformation, found that the root cause of most organizational problems stemmed from a lack of critical thinking. She suggests that leaders question their assumptions, particularly in high-stakes situations, reason through logic to avoid making false assumptions and seek diversity in thought and background from those with whom they collaborate to avoid groupthink. These are essential skills that have always been important in the business world and elsewhere but are even more needed with AI implementation to ensure that AI can deliver more accurate and reliable results, and that one can recognize when it goes astray. However, in her view, there is a big difference between “machine learning” and “machine thinking.” Considering AI as thinking machines, implies that we outsource our own critical thinking and problem-solving skills to a machine that is only replicating and reorganizing information it has gathered. Large language models are looking for existing patterns of information, perhaps even synthesizing them. But they cannot exercise judgment, no matter how nuanced or fast, their outputs are. While AI has clearly made significant advances over the last few years, it is still prone to several noteworthy limitations, such as the potential to create wrong information, generating biased outputs and demonstrating gaps in reasoning abilities. Even as AI gets more advanced, it absolutely requires the intervention of real people who can check for data inaccuracies, biases, logical inconsistencies and even proposals that go against ethical standards.
The Future: AI enhanced with Human Thought
The blend of human ingenuity and generative AI provides a powerful combination of expertise. While AI can perform some tasks exceptionally well, including data analysis and some forms of logic, it lacks the nuanced understanding and flexibility of human thinking. Listed are a few reasons why our critical thinking is so important as AI increasingly impacts our daily lives:
Contextual Understanding: AI lacks the ability to fully grasp the subtleties and nuances of human language and context. Critical thinking often includes understanding context, background information, and the broader implications of a situation, which can be challenging for AI to handle.
Creativity and Innovation: Critical thinking often involves creativity and thinking outside the box. While AI can generate solutions based on existing patterns and data, it struggles to generate truly novel ideas or approaches.
Ethical Considerations: Critical thinking involves ethical reasoning and moral judgment, which are complex and context-dependent. AI lacks human values and moral intuition, making it difficult for it to navigate ethical dilemmas effectively.
Emotional Intelligence: Critical thinking sometimes requires understanding and empathizing with different perspectives, emotions and experiences. AI lacks emotional intelligence and the ability to understand human emotions and motivations.
Adaptability: Critical thinking often requires adaptability and the ability to adjust reasoning strategies based on new information or changing circumstances. While AI can be trained on new data, it lacks the flexibility and adaptability of human thinking.
Conclusion
The application of AI tools requires an understanding of how the tool works and the content required. Moreover, to be useful, it also requires some level of critical thinking. While AI can be retrained on new data, it lacks human-type reasoning and the ability to update strategies in light of new information or shifting circumstances. When combined with human thinking, AI will enhance our productivity and add value to our work. Regardless of how an individual or a corporation is using AI, sharpening one’s own critical thinking will foster our ability to keep-up with the ongoing development of thinking machines